In this article I argue that labor practices that are detrimental to workers can be an antitrust problem even if the labor market is highly competitive. The reason is that even if a worker has many identical job offers ex ante, once the job match is formed, it is costly to dissolve. The desire to avoid these costs confers surplus that is specific to the match. This surplus will be divided via bilateral bargaining between the worker and the firm, and not in the context of the competitive labor market. Harmful labor practices may represent efforts by the firm to capture most or all of that surplus, and this can be thought of as an effort to monopolize that surplus with respect to that worker. One specific labor practice, namely imposing non-competes on workers, is particularly likely to be an antitrust problem, as it not only helps the firm capture the match-specific surplus, but also increases the surplus available to be captured by making it more difficult for the worker to re-access the competitive labor market.
In conventional antitrust analysis, there are certain conditions that must be met for a matter to be an antitrust problem. A merger is only likely to be a problem if the merging firms are close competitors to each other and there are not very many other close competitors. Coordination is only likely to be a problem if the coordinating firms collectively represent a large fraction of the sellers of the product. And conduct, including a variety of contractual restraints, is only likely to be a problem if the restraining firm has significant market power as that term is conventionally understood.3
Assuming, at least for the sake of argument, that this is true of output markets, it may appear that it must be true of input markets as well, including labor markets. Indeed, the textbook analysis of monopsony and oligopsony is very similar to that of monopoly and oligopoly,4 and this similarity is explicitly recognized in the DOJ/FTC Horizontal Merger Guidelines.5 It may therefore appear, and it is often argued, that labor market practices engaged in by firms can only be an antitrust problem in the presence of conventional labor market power, defined (roughly) as there being only one or a small number of possible employers for the type of labor in question.
The purpose of this article is to argue to the contrary. Specifically, I argue that firms can impose harmful conditions on workers even when workers have many roughly equivalent job offers to choose from.6 I further argue that this harm can reasonably be thought of as antitrust harm.7
The basic argument is as follows. Even if the labor market is very competitive ex ante (at the time of hiring), once the job match is formed, dissolving it is costly to both the worker and the firm, which means that preserving the match generates surplus, and this surplus must be divided between the worker and the firm. The more valuable the match (i.e. the more the worker prefers preserving the match to looking for another job, and the more the firm prefers preserving the match to looking for another worker), the more surplus there is to be divided.
This division will be determined via bilateral bargaining. This bargaining can be modeled using standard methods familiar from conventional antitrust analysis, specifically a model known as the Nash Bargaining model. In that model, the division of the surplus depends on the relative bargaining leverage between the worker and the firm, which loosely means that the less a party has to gain from reaching a deal, the more favorable the terms that party will receive; and on the relative bargaining power between the worker and the firm, which means that the stronger a party’s capabilities for capturing surplus, the larger the share of the surplus that party will receive. Having more bargaining leverage and more bargaining power are both beneficial to a party, but an important and under-emphasized result from the theory of Nash Bargaining is that a party that has all of the bargaining power captures all of the surplus, regardless of relative bargaining leverage. Specifically, if one party has all of the bargaining power, they can make a take-it-or-leave it offer to the other party that leaves that party no better off than they would be if the match was dissolved.
Whether a worker or a firm has more bargaining leverage is difficult to say, and there is no strong empirical evidence on the subject that I am aware of. However, there is some reason to suspect that workers often “need” a deal more than firms do, giving firms more relative bargaining leverage. With regard to bargaining power, matters are much clearer. Bargaining power is very likely to be held mostly by the firm, not the worker. Firms have myriad advantages in size, resources, and sophistication, and they can unilaterally set non-negotiable firm policies. An ordinary worker has little prospect of matching these advantages, and so will be at a major disadvantage in capturing the match-specific surplus.
Bargaining power as it appears in the Nash Bargaining model corresponds to power in the ordinary English sense of the word. While not without bound (the worker can still quit), the firm is able to use its advantages to acquire bargaining power, and to use that bargaining power to benefit itself at the expense of the worker. This can take the form of chiseling on wages and hours, or poor working conditions, or even abusive or degrading treatment.
The question is whether this power is market power in the antitrust sense. I argue that it is. As discussed above, the division of the match-specific surplus takes place outside the context of the competitive labor market, so the competitive labor market does not protect the worker from efforts by the firm to capture it.8 Practices that allow the firm to capture most or all of that surplus can be thought of as efforts to become a monopolist over that surplus with respect to that worker, which makes it an antitrust problem.9
The fact that harm from these practices might reasonably be thought of as antitrust harms does not necessarily mean that they should always be dealt with in the context of antitrust. In many cases, regulation by the Department of Labor or by OSHA may be more appropriate. Nor is it obvious which practices by a firm should or should not be regarded as antitrust violations. The purpose of this article is not to resolve these questions. The purpose is only to establish that there is a reasonable basis for considering these harms to be antitrust harms, and therefore to consider antitrust action as one possible avenue for addressing them.
There is one labor market practice that is particularly likely to be an antitrust problem, namely labor non-compete agreements. Unlike other labor practices, whose purpose is to affect the division of existing match-specific surplus, non-compete agreements have the effect of increasing the match-specific surplus. They do this by making it more difficult for the worker to re-access the competitive labor market, thereby degrading the worker’s prospects outside the match. When the worker’s outside option is worse, they value the match by more, increasing the amount of match-specific surplus available to be captured by the firm. Practices that distance workers from the opportunity to participate in competitive markets are quite clearly an antitrust problem.
This view of non-competes is at odds with the familiar argument that their voluntary nature is strong grounds for believing them to be both mutually beneficial and efficient.10 But this is only true if non-competes are truly voluntary, and if the agreed-upon compensation offered to the worker in exchange for accepting a non-compete is actually delivered. In a separate article (Balan, 2019), I argue that these conditions are often not satisfied.11 This argument is bolstered by a wealth of recent empirical evidence that, while somewhat mixed, largely finds non-competes to be harmful to workers.12 And if those conditions are not satisfied, then non-competes are not to be presumed beneficial to the worker, but are rather to be understood as a restraint imposed on workers without compensation, to the benefit of the firm and to the detriment of the worker
II. EX POST BARGAINING OVER MATCH-SPECIFIC SURPLUS
Suppose that a worker entering the labor market can choose between many jobs that ex ante appear to be identical. Once the worker chooses a job and a match is formed, that match is costly to dissolve, for both the worker and the firm. For the worker, the costs include the direct financial costs of a new job search, the lost income during the search (the damage from which is exacerbated by the fact that many workers have no financial cushion), and the fact that being fired is an emotionally traumatic experience for workers. In addition, the worker might need a recommendation from the firm to find another job, which they may not get if the match ends in acrimony. Finally, the worker may have signed a non-compete agreement or be subject to other restrictive covenants, which further increases the cost of leaving their job. For the firm, the costs include the direct recruiting and hiring costs to replace the worker, the indirect costs of being temporarily understaffed until a replacement is hired, and possible morale problems among remaining workers. These costs can be substantial.
Another way of saying that dissolving the match is costly is to say that preserving the match generates surplus arising from avoiding those costs. This surplus must somehow be divided between the worker and the firm. If labor contracts were complete and also fully and costlessly enforceable, then the division of this surplus would be determined within the context of the competitive labor market, and workers would receive competitive overall terms.
In reality, however, contracts are neither complete nor fully enforceable. This means that much of what happens between the worker and the firm is determined after the match has formed. There are things that the firm can do to benefit itself at the expense of the worker (chisel on wages and hours, poor working conditions, or even abusive or degrading treatment), and there are things that the worker can do to benefit themselves at the expense of the firm (shirking, theft, even sabotage). Which of these things will happen will be determined via bilateral bargaining between the worker and the firm, and not within the context of the competitive labor market.
This does not mean that the competitive labor market is irrelevant. A key concept in bargaining is that neither side can be forced to do worse than they would do if the match was dissolved (this is often referred to as their “outside option” or “disagreement payoff”). The worker will not agree to terms that are worse than being fired and having to look for another job, nor will the firm agree to terms that are worse than letting the worker quit and having to look for another worker. The more competitive the labor market, the better the worker’s outside option, and the better the terms the worker will receive. That is, a labor market characterized by conventional market power makes things worse for workers, for the conventional reasons. But even in a competitive labor market, a substantial amount of surplus will be divided via bilateral bargaining outside the context of the competitive labor market.
This kind of bilateral bargaining is standard in economics, including antitrust economics. The standard framework for studying it is a wellknown model called the Nash Bargaining model. In the remainder of this section, I summarize and present key results from this model.
A. Bargaining Leverage
As discussed above, the worker’s surplus from the match is the difference between what they receives if a deal is reached and what they receive if a deal is not reached, and similarly for the firm. The total surplus from the match is the sum of the worker’s surplus and the firm’s surplus.
Somewhat counter-intuitively, the party that contributes less to the total surplus has greater bargaining leverage. To see why, note that the surplus must be divided somehow. Suppose, for example, that the surplus is to be divided equally (though this is not necessary). A party that contributes little surplus shares little surplus with the other party (half of very little is also very little), but that party still receives half of the surplus contributed by the other party. An analogy would be a guest who brings a small side dish to a potluck dinner, but then eats the full meal like everybody else. Put another way, the party that contributes less to the total surplus has less to gain from a deal. That party “needs” a deal less, and so is able to bargain for better terms.
Whether workers or firms have greater relative bargaining leverage is difficult to say, and it may differ across employment matches and perhaps even over time within a match. While there is no direct empirical evidence that I am aware of, for the above reasons it appears that firms may often have greater relative bargaining leverage (i.e. having the match end is worse for the worker than it is for the firm).13
B. Bargaining Power
In Section II.A I assumed that the total surplus from reaching an agreement is divided equally between the parties. But this need not be the case. A given amount of surplus can be divided so that it goes entirely to one party, entirely to the other party, or anywhere in between. What share of the surplus a party can command is referred to as their bargaining power, which is distinct from the bargaining leverage described above. If one party has all of the bargaining power, then it will capture all of the surplus, and the other party will only receive value equal to their outside option, making them no better off than they would be if the match were dissolved. Any intermediate amount of bargaining power is also possible.
Bargaining power is an economic term of art, but it corresponds quite closely to the ordinary English usage of the word “power.” When there is a pool of surplus to be divided between a single worker and a large firm, who has the power to capture it? Is that division likely to be 50/50 (the worker and the firm share the surplus equally)? Or is it more likely to heavily favor the firm, say 90/10 or 95/5? The massive size, sophistication, and resources of the firm strongly suggest the latter, as does the fact that the firm unilaterally sets non-negotiable rules, policies, and employment practices that can be used to apply pressure to the worker. It simply strains credulity that ordinary individual workers can outperform large, heavily resourced firms in a competition to capture a pool of surplus. That is power, and it is the firms, not the workers, that have it.14
There is an additional point that is a standard result of Nash Bargaining, but that is not widely appreciated. If one party has (almost) all of the bargaining power, then it matters little who has more bargaining leverage. Recall that bargaining leverage is about the relative contributions of the two parties to the total pool of surplus. But if one party has all of the bargaining power, then this is moot, because that party receives all of the surplus, regardless of who contributed it. This will be made clearer in the next sub-section.
C. Numerical Examples
Define WD as the value that the worker receives if a deal is reached (the job match continues) and WND as the value that the worker receives if a deal is not reached (the job match is dissolved and the worker receives their outside option).15 The difference between these two (WD − WND) is the gain to the worker from reaching a deal. Note that (WD − WND) can be large because WD is large (getting a deal is very good), or because WND is small (not getting a deal is very bad), or some combination of the two. Similarly, define FD as the value that the firm receives if a deal is reached, FND as the value that the firm receives if a deal is not reached, and (FD − FND) as the gain to the firm from reaching a deal.
The total surplus TS from continuing the job match is
TS = (WD – WND) + (FD – FND),
where TS is the sum of the amount by which the worker is better off with a deal than without, plus the amount by which the firm is better off with a deal than without.
Each party will receive their outside option (WND for the worker and FND for the firm) plus some share of the match-specific surplus. For simplicity I assume that this surplus will be divided via a lump-sum payment P from the worker to the firm (this payment can be negative, which would mean a payment from the firm to the worker). It is important to note that this does not literally mean that the worker will hand money over to the firm, or vice-versa. Rather, the “payment” will take the form of one side or the other getting away with under-performing the terms of the original agreement, or with interpreting ambiguities in that agreement in a manner favorable to themselves. Workers may get away with a certain amount of shirking, and firms may get away with a certain amount of mistreatment of one kind or another.16
In our examples we will assume, as is common in antitrust economics, that P will be determined as predicted by the Nash Bargaining model. There are five inputs into this model: WD, WND, FD, FND, and a “bargaining power” parameter α that governs the share of the surplus that is kept by the worker. According to the model, the equilibrium P will be the one that maximizes the following expression:
((WD – P) – WND) ∝ ((FD + P) – FND) 1–∝).
Less technically, the P that comes out of the Nash Bargaining Model is the one that causes the worker to receive their outside option plus surplus equal to α TS, and the firm to receive its outside option plus surplus equal to (1 − α)TS.
Now consider the following examples. In each example, TS = 200.
Example 1: (WD − WND) = 100, (FD − FND) = 100, and α = ½.
The two parties are identically positioned, so we would expect P = 0 to be the answer. This is indeed the case. The worker and the firm have equal bargaining leverage; they each prefer a deal to no deal by 100, so they each contribute 100 to the TS = 200. They also have equal bargaining power, because α = ½ and (1 − α) = ½, so they each are to receive their outside option plus 100 (half of the TS) net of P. P = 0 is the P that accomplishes this, as this is what they each already receive gross of P.
Example 2: (WD − WND) = 150, (FD − FND) = 50, and α = ½.
Now the firm has more relative bargaining leverage than in Example #1, because the worker prefers a deal to no deal by 150, and the firm prefers a deal to no deal by only 50, meaning that the worker contributes more than half of TS. But as in Example #1, the bargaining power is equal (α = ½), so the worker and the firm will each receive their outside option plus half of TS (i.e. their outside option plus 100) net of P. Since the worker receives their outside option plus 150 gross of P, and the firm receives its outside option plus 50 gross of P, a payment of P = 50 is required. If the 150 and the 50 were reversed, then P would be –50, and the firm would be paying the worker.
Example 3: (WD − WND) = 100, (FD − FND) = 100, and α = 0.
Now the two parties have the same bargaining leverage (as in Example #1), with each preferring a deal to no deal by 100, and so each contributing 100 to TS. But now the firm has all of the bargaining power, which means that the worker will receive no surplus net of P, only receiving their outside option. The firm will receive its outside option plus 200 (all of TS) net of P. Since they each receive their outside option plus 100 gross of P, a payment of P = 100 is required. If we replaced α = 0 with α = 1, P would be –100, the firm would receive its outside option, and the worker would receive their outside option plus 200.
Example 4: (WD − WND) = 0, (FD − FND) = 200, and α = 0.
Now the worker has all of the bargaining leverage (they are indifferent between a deal and no deal, but the firm prefers a deal to no deal by 200 and so contributes all of the TS). But as in Example #3, the firm has all of the bargaining power. In this case, the worker still receives no surplus net of P, only receiving their outside option, and the firm still receives its outside option plus 200 (all of TS) net of P. Since the worker receives no surplus beyond its outside option gross of P, the payment will be P = 0. If we replaced α = 0 with α = 1, P would be –200, the firm would receive its outside option, and the worker would receive their outside option plus 200.
This last example is somewhat subtle but very important. If the firm has most or all of the bargaining power, then it will receive most or all of the bargaining surplus, regardless of bargaining leverage. When one side can capture all of the surplus from a deal, it does not matter who contributed how much to that surplus.
These examples are highly stylized, and they omit many important factors. However, they capture an essential point, namely that when the firm has a lot of bargaining leverage, and (more importantly) when the firm has almost all of the bargaining power, it is possible for workers to be harmed by firms ex post even when they participate in a highly competitive labor market ex ante.
III. IS THIS AN ANTITRUST PROBLEM?
The existence of costs to dissolving a job match creates match-specific surplus, and that surplus must be divided somehow between the worker and the firm. Labor market practices that firms engage in to the detriment of workers can be understood as efforts to capture that surplus. There are many such practices, including chiseling on wages and hours, poor working conditions, or even abusive or degrading treatment.17
Some of these practices may be actionable under labor law, but the question for this article is whether they can be considered antitrust violations, even if the labor market is highly competitive ex ante.18 I argue that they can. The key fact is that the surplus that these practices are intended to capture are specific to the job match, and so by definition the harm suffered by the worker from these practices cannot be ameliorated by labor market competition. These are practices of a single firm against the worker. When the firm has disproportionate bargaining leverage or (more importantly) most or all of the bargaining power, it can reasonably be regarded as a monopolist with respect to that worker over the match-specific surplus.
It might be argued that these practices represent a permissible exercise of existing market power, and not an impermissible acquisition of market power. Setting aside the question of whether exercising existing market power is in fact always permissible, I believe that the practices represent the acquisition of market power, and not the exercise of it. It is true that the practices do not create the surplus, which exists exogenously by virtue of the value of the match. But that surplus does not start out belonging to the firm. The surplus exists and it must be divided somehow, and the labor practices are the means by which that division comes out to the benefit of the firm at the expense of the worker. The exercise of bargaining power is the exercise of market power.
The above is not a worked-out apparatus for treating labor practices as antitrust violations. There are many questions that are beyond the scope of this article, including which practices are harmful at all, and of those which are best dealt with through labor law rather than antitrust. There is also the conceptual question of how much of the match-specific surplus firms should be allowed to try to capture. Should they be allowed to try to capture half of it? Would an antitrust case hinge on what fraction of the surplus the firm refrained from trying to capture?
These are difficult questions, and it is unclear whether it is possible to build a workable regime for challenging harmful labor market practices as antitrust violations. It may be or it may not be. But the fundamental point remains. These practices represent firms trying to become monopolies with respect to their workers regarding the match-specific surplus. This is the exercise of market power.
IV. LABOR NON-COMPETE AGREEMENTS
Another labor practice that firms sometimes engage in is to impose non-compete agreements on their workers. Non-competes are fundamentally different from the labor practices discussed in Section III above. Those practices represent attempts by the firm to capture a fixed quantity of match-specific surplus. In contrast, non-compete agreements increase the amount of surplus available to be captured. A non-compete agreement denies the worker access to the full benefits of the competitive labor market, thereby degrading that worker’s outside option. The worker now has more to gain from the match, increasing the total surplus arising from the match, and the firm can use its superior bargaining power to capture most or all of that additional surplus as well.19
For this reason, the argument for treating non-compete agreements as an antitrust problem is even stronger than the argument discussed above for treating other labor practices as antitrust problems. A practice that denies the worker the ability to re-access the full benefits of the competitive labor market appears to fall quite squarely within the domain of antitrust, especially when combined with the firm’s ability to use its bargaining power to capture the resulting increased surplus.
As discussed above, if contracts were complete and fully and costlessly enforceable, all terms of the labor contract would be determined in the context of the competitive labor market, and hence restraints such as non-competes would not be an antitrust problem (assuming that they also did not harm third parties). This is closely related to a standard defense of non-competes, namely that workers would not agree to them unless they receive compensation that they value at least as much as they dislike the restraint. In a separate article (Balan, 2019), I argue that this is often not the case, and that in fact non-competes are a means of extracting value from workers without having to compensate them for it.20
There is reasonable consensus that conventional labor market power can exist when there are only one or a few employers that hire a particular type of worker, and that antitrust is applicable to those situations. Some hold the view that the existence of such market power is a necessary condition for antitrust to apply to labor markets, meaning that when there are many employers who hire a particular type of worker, any problems that may arise from the conduct of an individual firm cannot be antitrust problems.
The purpose of this article is to argue against this view. Even with an ex ante competitive labor market, once a job match is formed, dissolving it is costly to one or both parties, meaning that there is often substantial economic surplus associated with continuing it. The division of this surplus will be determined via bilateral bargaining between the two parties, and not within the context of the competitive labor market.
Firms often have major advantages over workers in capturing that surplus. They often have more relative bargaining leverage, as workers may “need” the match more than they do. More importantly, firms almost certainly have much more bargaining power. Given the massive asymmetry of resources, sophistication, and agenda-setting power between an individual worker and a large firm, it strains credulity that the firm would not have a massive advantage allowing it to out-compete the worker in any contest to capture it. This gives the firm power over the worker, in the ordinary English meaning of the word, in the formal meaning of the word in the context of the Nash Bargaining model, and in the antitrust sense; certain labor market practices represent an attempt to become a monopolist with respect to the worker over that match-specific surplus.
The case for treating non-compete agreements as an antitrust problem is even stronger. Firms imposing non-competes on workers is not only a means of capturing an existing quantity of surplus, it is a way of increasing that surplus by denying the worker the ability to fully access the competitive labor market (degrading the worker’s outside option), and then using its power to capture that additional surplus as well.
1 David Balan is an employee of the Federal Trade Commission. The views expressed in this paper are solely those of the author.
2 I thank Keith Brand, Wally Mullin, and Jeremy Sandford for helpful comments.
3 This is a crude one-paragraph summary of “conventional antitrust” for the purpose of fixing ideas, and should not be taken literally. It should also not be taken as an endorsement of what is currently conventional antitrust, nor as an endorsement of any particular alternative thereto.
4 In a perfectly competitive labor market, a firm that reduced wages by even a penny would not be able to attract any workers, so it has no choice but to pay the competitive wage. A firm with market power, in contrast, faces a tradeoff. It can raise wages to attract more workers, which it will want to do as long as the profits derived from each incremental worker exceeds the wage. But if it increases the wage for incremental workers, it will (in most instances) have to increase wages for the original workers as well. This limits the incentive to increase wages, and so wages (and also employment) are lower than they would be under competition. How much lower depends on the magnitude of the firm’s market power. This is very similar to the reasoning behind market power in output markets.
5 FTC/DOJ Horizontal Merger Guidelines, 2010, Section 12.
6 Recent research provides some evidence that conventional market power is sometimes present even when it appears not to be. For example, there are many fast food chains, and the jobs that they offer appear to be quite similar, and yet there is some evidence that an individual fast food chain can exercise some labor market power. While this is important research, for the purposes of this article I assume that it does not apply. That is, I assume that labor markets are highly competitive in the sense that workers really do have a large number of ex ante equivalent job offers to choose from.
7 I do not offer an opinion about whether these labor market harms are actionable under the antitrust laws as they currently exist and as they are currently interpreted by the antitrust agencies and by the courts. Nor do I offer an opinion about what priority these harms, if they exist, should receive compared to other antitrust agency enforcement priorities. My argument is only that, as a conceptual matter, these labor market harms, when properly understood, can reasonably be regarded as antitrust problems.
8 The competitive labor market still benefits the worker by strengthening their outside option; a labor market characterized by conventional market power would make matters worse. But it does not help the worker in its bargaining with the firm over the match-specific surplus.
9 Recent research by Stansbury & Summers makes the related claim that important U.S. economic trends are the result of declining worker power. (See Anna Stansbury & Lawrence Summers, “The Declining Worker Power Hypothesis: An Explanation for the Recent Evolution of the American Economy,” NBER Working Paper 27193, May 2020.) However, in that research the term “worker power” refers to the ability of workers to capture a share of a firm’s profits or rents. This is different from the bargaining power discussed in this article, which is confined to a worker’s ability (or inability) to capture match-specific job surplus. This distinction is important for the purposes of this article, because it is the match-specific nature of this surplus (i.e. the fact that it is divided via bilateral bargaining outside the context of the competitive labor market) that makes it specifically an antitrust problem.
10 A non-compete could mutually beneficial but not efficient if third parties are harmed.
11 David J. Balan, “Are Labor Non-Compete Agreements Efficient?” mimeo, 2019.
12 See Evan Starr, “The Use, Abuse, and Enforceability of Non-Compete and No-Poach Agreements: A Brief Review of the Theory, Evidence, and Recent Reform Efforts,” Economic Innovation Group Issue Brief, February 2019. See also Evan Starr & Michael Lipsitz, “Low-Wage Workers and the Enforceability of Non-Compete Agreements,” mimeo, 2020.
13 For simplicity, I assume that a dollar is equally important to the worker and to the firm. But in fact, a dollar likely means more to the worker than it does to the firm. So even if a vacancy costs the firm more money than unemployment costs the worker, it may still be the case that dissolving the match is more painful for the worker than it is for the firm, which would tend to increase the relative bargaining leverage of the firm at the expense of the worker.
14 There may be exceptions to this, for example a CEO negotiating with representatives of a firm’s board may have a significant amount of bargaining power. (See Omesh Kini, Ryan Williams & David Yin, “CEO Non-Compete Agreements, Job Risk, and Compensation” Review of Financial Studies, 2020.) But for ordinary workers at large firms, it is difficult to imagine that they can successfully compete in a contest to capture a quantity of surplus.
15 The outside option is what the worker would receive if the match were dissolved and they got another job. But if the worker got another job, there would be match-specific surplus from that job, which would also be bargained over. If all jobs were literally identical, there would be an infinite recursion of this problem, which would make it impossible to pin down the outside option in the Nash Bargaining model. There are a number of assumptions that one could make to get around this problem. One simple option is to assume some ex ante unobserved heterogeneity among firms, such that a worker who got a particularly bad draw would (at least in expectation) get a better draw next time. The more likely that is, the better the outside option.
16 In footnote #13 above, I point out that a worker and a firm likely do not have the same value of a dollar, with the worker likely suffering greater harm from the loss of a dollar than does the firm. This, combined with the fact that in the real world the “payment” often takes the form of acquiescing to undesirable behavior by the other party rather than transfers of money (the main exception being wage and hours violations, which can be thought of as money transfers from the worker to the firm), complicate the model because the P paid by one party is no longer necessarily equal to the P received by the other party
17 Abusive or degrading treatment presumably does not benefit the firm. But it may be something that managers have a preference for, and the firm may have limited incentive or ability to prevent the managers from engaging in it. So in a sense the ability to indulge that preference can represent the manager capturing part of the surplus at the expense of both the worker and the firm.
18 Recall that the maintained assumption throughout this article is that the labor market is highly competitive ex ante. If this is not the case, then these practices can be antitrust violations for conventional reasons.
19 It is a property of the Nash Bargaining model that an increase in the match-specific surplus arising from the worker valuing the job more is no different from an increase in the match-specific surplus arising from the worker having a worse outside option. Both contribute equally to the total surplus from the match, and both will be captured by the firm if the firm has most or all of the bargaining power.
20 See David J. Balan, “Are Labor Non-Compete Agreements Efficient?” mimeo, 2019. A sketch of the argument in that piece is as follows. First, the worker may not even know that the job comes with a non-compete until after they have begun work and turned down their other job offers. Second, the worker may not fully understand what the non-compete means, or it may be so abstract that it is not salient to them. Third, just because compensation was promised in exchange for agreeing to the non-compete does not mean that the compensation will be delivered, and if it is not delivered, the non-compete itself closes off the primary remedy, namely the ability to quit and get another job. Fourth, in many instances, there is not even the possibility of negotiating over the non-compete. In many cases it is simply firm policy, without exceptions. And once requiring a non-compete becomes standard in this way, the practice is likely to be very difficult for competition to dislodge.